Variant analysis
Variant analysis groups cases by the unique sequence of activities they followed and ranks variants by frequency and cost. The top 3-5 variants typically account for 70-90% of cases; the long tail of rare variants is often where rework, errors, and process exceptions hide.
Variant analysis is process mining's most visceral output: a list of distinct paths sorted by how many cases followed each. The first variant is usually the happy path; variants 2-5 are often deliberate alternatives; variants 6+ are typically a mix of exception paths and process anomalies. The pragmatic use: look at variants 6-20 and ask 'why does this variant exist?'. Common answers include manual workarounds for system limitations, special handling for specific customer types that wasn't documented, and process drift that nobody noticed. Each variant is a candidate for either standardisation (codify into the main process) or elimination (fix whatever forced the workaround).
Discussed in our use-cases
ICP-targeted pages where variant analysis is part of the framing.
Related terms
- Process discovery
Process discovery is the process-mining technique of constructing a process model (typically BPMN or a Petri net) from an event log without prior knowledge of the intended process.
- Process conformance
Process conformance is the analysis of how well an observed event log matches a reference process model — measuring deviations, missing steps, and extra steps.
- Happy path
The happy path is the variant of a process in which everything goes as intended — no exceptions, no rework, no manual interventions.